The five types of paired indicators

“The U.S. government is a structure designed (in part) to make it difficult to get things done.” That’s one of things I took from Robert Caro’s The Years of Lyndon Johnson series. The third volume–Master of the Senate—which details Johnson’s stratagems and manoeuvring within the Senate made this particularly apparent. Having grown up in the South West of the United Kingdom I knew nothing about such things, so reading those books was my first real exposure to the system of checks and balances that is U.S. governance.

Naturally, it raised some questions. The first—and perhaps the most naive—was, “Why make it so difficult for the people in power to make decisions and take action?” The answer, as I discovered upon reflection, is quite simple; the system of checks and balances is a hedge against risk. Any form of democracy sacrifices agility and velocity in order to ensure that power cannot be consolidated in—and thus, abused by—a single entity. 

This is reversed in other types of government like, for example, a dictatorship. One of the advantages of such supreme authority is that decisions can be made and actions can be taken with astonishing speed. Consider the Third Reich. If someone wanted to bring in a discriminatory law or put pressure on races they considered inferior, what needed to happen? They didn’t need to draft bills, get them approved by government departments, and have them voted on by the Reichstag. Nope. All they needed was Hitler to approve it. And because of the power Hitler had consolidated in himself, anything he approved or wanted done was undertaken immediately.

Of course, this dynamism is probably the only advantage to rule which places all power in one person’s hand. There aren’t any others. So it’s better to have institutions and processes which spread power out and minimise the chance of it being abused, to have mechanisms that mitigate against the dominance of a single entity, even at the cost of decision making speed.

This method of mitigating risk with a system of checks and balances is also applicable to business management. In High Output Management, Andy Grove talks about “pairing indicators”:

“Indicators tend to direct your attention toward what they are monitoring. It is like riding a bicycle: you will probably steer it where you are looking. If, for example, you start measuring your inventory levels carefully, you are likely to take action to drive your inventory levels down, which is good up to a point. But your inventories could become so lean that you can’t react to changes in demand without creating shortages. So because indicators direct one’s activities, you should guard against overreacting. This you can do by pairing indicators, so that together both effect and counter-effect are measured.”

Now, I’m not a CEO or a manager. I don’t have inventory to monitor or administrative tasks to oversee. But it did get me thinking. “How can I apply this idea of pairing indicators to my own life?” The actual answers are numerous and probably quite yawn-inducing, so I’ll save you the specifics of applying this idea in each domain of my own life. But what is more interesting is the types of pairs you can opt for. One example comes from a16z co-founder Marc Andreessen:

“The problem with arbitrary metrics in complex situations – they tend to backfire. Give emergency services drivers rigid response time metrics, and they’ll tend to stay close in to urban centers. Surprise!

Andy Grove had the answer: For every metric, there should another “paired” metric that addresses adverse consequences of the first metric. Many companies and especially governments violate this principle continuously, and are startled by the result – every time.

A large amount of the current Western (and Asian) obsession with quantitative test scores in education has this problem. Are we trying to train students to be test takers, or flexible/capable people who can thrive in complex/uncertain situations? If you’re worried about robots eating all the jobs, maybe we should stop programming students as if they’re robots.”

Andreessen’s observation is that we can pair a metric derived from a legible task with one derived from an illegible task to get a more complete picture of a person’s competence in complex environments. But legible-illegible isn’t the only way to pair metrics. There are others.

The most obvious is short-term-long-term. Pair one metric which measures short-term performance with one which emphasises long-term performance or trajectory. One example of such a pairing is Eric Ries creation of the Long Term Stock Exchange. He’s attempting to partner metrics that encourage short-sighted action with those that reward behaviours that promote sustainable value creation far into the future.

Another type of pairing is quantitative-qualitative. For example, an individual could monitor growth in income alongside stress levels and contentment. This would mitigate the tendency for income growth to take over a person’s life.

Similarly, an organisation or person could pair process-outcome metrics. Using my own work as an example, I could instigate a system which tracks the number of words I put out and monitor the hours I spend doing deep work or immersed in a state of flow.

Yet another metric pairing comes from Gabriel Weinberg’s Traction. That book contains something called the 50 percent rule: “…spend 50% of your time on product and 50% on traction.” It’s spending half your time on your market offering, and half your time on your position in a market. The metric pairing that arises from this? Internal-external. Another example of this type of pairing is a metric that measures how those within the company perceive it and a metric that measures how those outside the company perceive it. 

There may well be other types of paired indicators, but those are the ones that stand out to me. Legible—illegible. Short-term—long-term. Quantitative—qualitative. Process—outcome. Internal—external. All these pairs ensure that, in the words of Grove, “both effect and counter-effect are measured.”